Smartsolar 100/20 does not recognize PV voltage

Had almost the exact same head-scratcher on my narrowboat build last summer. Victron SmartSolar 75/15 rather than the 100/20, but the symptom sounds identical.

Panels were reading a healthy open-circuit voltage at the cable ends — something like 38V on my two Renogy 100W panels in series — but the moment I connected to the MPPT it dropped to battery voltage. Completely baffling.

Turned out to be a failed bypass diode inside one of the panels. Weird thing is it only showed up under load. Open-circuit the panel looks fine, but once the controller tries to pull current the voltage collapses. The dodgy diode was basically short-circuiting half the panel string.

Worth checking a few things:

• Test each panel individually with a proper load (not just open-circuit volts)
• Check your MC4 connectors — a corroded or poorly crimped connector can do this too
• Have a look in the VictronConnect app under the error log, it might throw up an error code that narrows it down

Also worth confirming your VOC in series isn't creeping above the 100V input limit on that controller — 2x170W could be pushing it depending on the panels.

Has anyone else seen this? Curious whether it's more common with certain panel brands. I've heard Fogstar's newer panels have solid QC but haven't tried them in a proper series string yet.

@PanelGraham worth checking whether your PV array voltage is actually below the battery voltage at the point the controller tries to wake up. The SmartSolar needs PV to exceed battery by a certain threshold before it'll engage — something like 5V over battery voltage from memory.

Had this exact issue on my static caravan setup. Panels were partially shaded first thing in the morning, so the controller just sat there looking confused until mid-morning when irradiance picked up properly.

Also worth ruling out:

• Corroded MC4 connectors (surprisingly common, even on "weatherproof" ones)
• Wrong panel wiring — series vs parallel matters for hitting that minimum threshold
• Faulty PV fuse inline

The VictronConnect app will show you live PV voltage readings which makes diagnosis much easier. If it's reading zero or something nonsensical, that points more toward a wiring/connector fault than the controller itself.

@PanelGraham @RetiredPlumber — worth adding: the 100/20 has a minimum PV voltage threshold of roughly battery voltage + 5V before it'll even attempt to start. If your panels are wired in a configuration that's marginal at dawn or under heavy cloud, the controller simply won't wake up.

Also check the VE.Direct history in the Victron Connect app — specifically "minimum PV voltage recorded." I've caught wiring faults on my shepherds hut system that weren't obvious until I looked at historical minimums.

One other thing people miss: reverse polarity on PV terminals won't blow the 100/20 immediately, but it can cause exactly this symptom. Measure PV voltage at the controller terminals, not just at the panel junction box. Voltage drop across dodgy MC4 connectors is surprisingly common on narrowboat builds where connectors get damp.

Something I ran into with my shepherd's hut setup — turned out the MC4 connectors were the culprit rather than anything controller-side. One had a slightly recessed pin that looked fine visually but was creating enough resistance to drag the voltage reading down into a weird limbo state. The MPPT just sat there confused.

Worth doing a proper voltage measurement at the controller's PV terminals directly rather than trusting what VictronConnect is reporting — there can be a meaningful difference if your cable run has dodgy joins or undersized wire. Mine was showing ~18V in the app but I was getting 21V at the panel and only 16.8V actually arriving at the controller terminals. That was enough to cause exactly the symptoms you're describing.

@MarshLover MC4s are sneaky little devils for this — had exactly that on my first LiFePO4 build. Looked perfect from the outside, measured fine with a multimeter on open circuit, but under any real load the connection was dropping enough voltage to confuse the controller completely.

The test that caught mine: I put a small load across the panel output before the MPPT and watched the voltage on the Victron app. Soon as current actually flowed, the reading collapsed. Classic sign of a dodgy crimp rather than anything wrong with the controller itself.

Worth investing in a proper MC4 crimping tool — I borrowed one initially and eventually bought the Amprobe version. Cheaper connectors from Amazon are also notorious for this. Genuine Stäubli or at least reputable branded ones made a world of difference on my setup.